Databricks Unveils Four Key Updates to Enhance Enterprise Control Over AI Agent Development
These tools aim to address the challenges enterprises face when deploying AI agents for high-value use cases
On Tuesday, Databricks unveiled a suite of new tools aimed at helping enterprises scale AI agents from pilot phases to full production, while ensuring robust governance, monitoring, and seamless integration. The tools—Mosaic AI Gateway, AI/BI Genie Conversational API, Agent Evaluation Review App, and Batch AI—are currently available in public preview.
These tools aim to address the challenges enterprises face when deploying AI agents for high-value use cases. According to Databricks, 85% of global enterprises are using generative AI, but many models fall short of delivering accurate, well-governed outputs due to a lack of awareness of enterprise data.
“Enterprises often struggle to deploy AI agents effectively due to concerns over accuracy, governance, and security. For these organizations, the real challenge lies not just in technology, but in building the confidence to fully leverage the data intelligence benefits of generative AI,” said Craig Wiley, senior director of AI/ML product at Databricks. “These new tools directly address these issues, enabling businesses to scale beyond pilots and deploy AI agents with confidence.”
The Mosaic AI Gateway provides centralized governance by integrating both open-source and commercial AI models into a single platform. It supports custom large language models (LLMs) and ensures consistent governance, monitoring, and integration across these models.
Meanwhile, the AI/BI Genie Conversational API enables developers to embed AI-powered chatbots into custom applications and productivity tools, such as Microsoft Teams, SharePoint, and Slack. The API ensures continuity in conversations by retaining context, allowing for follow-up queries without losing track of previous discussions.
The Agent Evaluation Review App simplifies human-in-the-loop workflows by allowing domain experts to provide structured feedback, send traces for labeling, and customize evaluation criteria. This eliminates the need for spreadsheets or custom applications.
Finally, the provision-less batch inference tool enables enterprises to run batch inference with just a single SQL query, fully integrating with Mosaic AI without requiring any infrastructure provisioning.
Earlier this year, Databricks secured $15.3 billion in financing, bringing its valuation to $62 billion. This round was led by JPMorgan Chase, Barclays, Citi, Goldman Sachs, and Morgan Stanley and included $10 billion in Series J equity funding along with $5.25 billion in debt financing.
